An Improved Particle Swarm Optimizer with Shuffled Sub-swarms and Its Application in Soft-sensor of Gasoline Endpoint

摘要

This paper proposes a shuffled sub-swarms particle optimizer algorithm (SSPSO) to enhance the diversity of particles in the swarm to improve the performance of PSO. SSPSO is tested with a series of benchmark functions and compared with other version PSO algorithms. Experimental results show that SSPSO improves the search performance on the benchmark functions significantly. Furthermore, SSPSO is used to train NN to construct an artificial neural network SSPSONN. Then SSPSONN is applied to construct a soft-sensor of gasoline endpoint and compared with actual industrial data, the results show that the constructed soft-sensor is feasible and effective.